Method for identifying real estate secondary market structure based on fractal structure
Through the head-to-tail segmentation and characteristic price model method based on fractal structure, the real estate sub-market was identified, which solved the problem that the existing technology could not accurately identify the complexity of the spatial distribution of housing prices and the relationship between market nesting, and achieved an accurate understanding of the real estate market structure and Policy support.
Patent Information
- Application Number
- CN202411025407.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-30
- Publication Date
- 2024-11-12
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
It is difficult for the existing technology to effectively identify the structure of the real estate sub-market, and traditional methods cannot accurately demonstrate the complexity of the spatial distribution of housing prices and the market nested relationship, and the cost across systems is relatively high.
A fractal structure-based method is adopted, combining the head-to-tail segmentation and characteristic price model, and identify the real estate sub-market from the bottom up, analyze the housing price structure through the fractal perspective, and establish a new understanding of the real estate market economy.
It has achieved accurate identification of the nested structure of the real estate market space, provided a new perspective and method for policy implementation, and improved the accuracy of real estate valuation and understanding of market structure.
Smart Images

Figure SMS_1 
Figure SMS_5 
Figure FDA0004968392440000011
Abstract
Description
Technical Field
[0001] The invention relates to the field of real estate market structure research, and in particular to a method for identifying real estate secondary market structure based on fractal structure. Background Art
[0002] As the scale of the real estate market continues to expand, the difference in real estate prices between different cities has gradually widened, and the national real estate market has shown obvious differentiation, with significant regional differences. To explore a new development model for the housing economy, it is necessary to identify the real estate market structure more clearly and accurately.
[0003] The spatial heterogeneity of housing prices is a key issue in the study of real estate market structure. The urban housing economy is affected by various spatial factors and shows obvious spatial heterogeneity, which in turn shows the fractal characteristics of a complex spatial system. Therefore, in the study of the real estate market structure, traditional qualitative research and quantitative analysis based on geometric measurement are no longer accurate and effective for complex systems such as cities, and fractal theory provides us with an opportunity to understand the emergence of the internal structure of the real estate market. Studying the real estate market based on fractal structure can not only improve the accuracy of real estate valuation, but also help to correctly understand the natural housing economic structure of the city, which is conducive to formulating real estate policies that are more in line with the laws of reality.
[0004] Therefore, an effective method based on fractal structure to more accurately identify the secondary real estate market is urgently needed. At present, the classification of the secondary real estate market is mostly based on the multi-dimensional attributes of the characteristic price model, and the cost of cross-system comparison and method migration is high; and the existing real estate market classification method cannot effectively show the complexity of the spatial distribution of housing prices and the nested relationship of the real estate market. Fractal structure identification is only based on real estate prices, the required data conditions are loose, and it is meaningful to understand the combination of fractal theory and real estate economic structure. However, there is currently no complete method for identifying the secondary real estate market based on fractal structure. Summary of the invention
[0005] In view of the above problems existing in the prior art, the present invention provides a method for identifying the structure of the real estate secondary market based on fractal structure. Based on the fractal structure of the real estate market price, the method combines the head-tail segmentation and the characteristic price model to identify the real estate secondary market from the bottom up. By identifying the structure of real estate prices themselves from a fractal perspective to establish a new understanding of the real estate market economy, a new perspective and method can be provided for understanding the spatial nested structure of the real estate market and the implementation of policies.
[0006] The technical solution of the present invention is as follows:
[0007] A method for identifying real estate secondary market structure based on fractal structure comprises the following steps:
[0008] S1. Collect spatial vector data of housing prices in the main built-up areas of the research city, define them as housing price sample points, and obtain a set HP; the housing price sample points include: housing unit price information, housing structure information and housing latitude and longitude information; the housing structure information includes: area, number of bedrooms, number of bathrooms and floors;
[0009] S2. Execute the head-to-tail segmentation loop on HP. The specific steps are as follows:
[0010] S2-1. Set the values of the main parameters: number of divisions n, head-to-tail division ratio p, housing price sample point contact distance threshold r, housing price point sample size s;
[0011] S2-2, let the number of divisions n = n + 1, and calculate the average value m of the current HP; classify the housing price sample points greater than or equal to m as the head housing price points, and obtain the set H n ; Classify the housing price sample points less than m as tail housing price points, and get the set T n ;
[0012] S2-3. Calculate H n The sample proportion p′ is:
[0013]
[0014] S2-4. If p′ is less than or equal to p, then T n Defined as the initial secondary market HT for the current split number n , and execute S2-5 to S2-7; otherwise, HP is defined as the initial secondary market HT of the current split number n , and execute S2-5, S2-6 and S2-8;
[0015] S2-5. Initial secondary market HT n The housing price sample points are networked according to the house longitude and latitude information, and the weight of the network edge is the Euclidean distance between the housing price sample points;
[0016] S2-6. Delete the network edges with weight greater than r, calculate the size of each subgraph in the network, delete the housing price sample points with subgraph size less than 3, and obtain the new secondary market HT n ′;
[0017] S2-7. Calculate HT n ′, the number of sample points of housing prices s′; if s′ is greater than or equal to s, then HT n ′ is defined as the natural secondary market NHT n , execute S2-9; otherwise, HT n Add the head H n As the new HT n , execute S2-5, S2-6 and S2-8;
[0018] S2-8. Calculate HT n ′, the number of sample points of housing prices s′; if s′ is greater than or equal to s, then HT n ′ is defined as the natural secondary market NHT n , execute S2-10; otherwise, let n = n-1, and then divide T with n times of division n With HT n Add together to get the new HT n , then re-execute S2-5, S2-6 and S2-8 until s′ is greater than or equal to s, then execute S2-10;
[0019] S2-9, NHT n As the new HP, return to S2-2 and enter the next cycle;
[0020] S2-10, the cycle ends and all natural secondary market NHT is output n ;
[0021] S3. The obtained natural market NHT n Establish a hedonic price model, using variables including: housing structure information, location characteristics and facility distance; use multiple linear regression to obtain different natural market NHT n The fitting results are as follows: According to the adjacent natural market after merging relative to the single natural market NHT n The overall R 2 Increment size to obtain the final real estate sub-market segmentation results.
[0022] 4. Further, the specific steps of S3 are as follows:
[0023] S3-1. Natural Market NHT n The number of is denoted as k;
[0024] S3-2, starting from the two-by-two merger, select 2 to k adjacent and continuous natural markets NHT n , merge the corresponding housing price sample points and establish a characteristic price model for multiple linear regression fitting to obtain the corresponding regression fit goodness of fit
[0025] S3-3. Calculate the relative value of each natural market combination to the NHT alone. n The increase in the goodness of fit of the multiple linear regression Select the maximum increment The natural market merger situation, the merged secondary market output as NS n , and obtain the final real estate secondary market segmentation results.
[0026] Furthermore, the number of segmentation times n=0, the head-tail segmentation ratio p=0.5, the housing price point connection distance threshold r=800, and the housing price point sample size s=200.
[0027] The beneficial technical effects of the present invention are:
[0028] (1) Based on fractal theory, the present invention proposes an improved head-to-tail segmentation and merging method, analyzes the intrinsic structural characteristics of housing prices into housing sub-markets, and establishes a universal housing market segmentation method.
[0029] (2) The housing market segmentation method proposed in the present invention is based on the hierarchical structure and self-similarity characteristics of real estate market prices, and can conduct cross-system comparisons of different cities and provide new understandings of the housing market structures of multiple cities. BRIEF DESCRIPTION OF THE DRAWINGS
[0030] Figure 1 To improve the head-tail split loop implementation flow chart;
[0031] Figure 2 This is a sub-market structure diagram for the Yangtze River Delta region. DETAILED DESCRIPTION
[0032] The present invention is described in detail below in conjunction with the accompanying drawings and embodiments. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0033] The embodiment adopts the present invention to identify the structure of real estate sub-markets in the 26 major cities in the Yangtze River Delta region. The specific steps are as follows:
[0034] S1. Collect housing price data in the Yangtze River Delta region by crawling second-hand housing transaction websites such as Anjuke and Beike (or other data acquisition methods), including house unit price information, house structural attributes such as area, number of bedrooms, number of bathrooms, and floors to facilitate the subsequent establishment of a characteristic price model, as well as longitude and latitude information for geocoding and processing into shapefile spatial data.
[0035] S2. Correct the housing price data to obtain data in the WGS1984 coordinate system, and use the spatial location filtering tool in ArcGIS Pro to filter out the housing price data using the administrative division data of 26 cities in the Yangtze River Delta (also in the WGS1984 coordinate system) to eliminate the location errors caused by geocoding. The filtered housing price data is defined as the housing price sample points, and the set of housing price sample points is HP.
[0036] S3. Perform improved head-to-tail segmentation on the HP of 26 cities in turn. Determine the main parameters of the head-to-tail segmentation process: head ratio p = 0.5, spatial distance r = 800, scale s = 200. The specific steps of head-to-tail segmentation for each city are as follows: Figure 1 As shown:
[0037] S3-1, input the HP of the city, and record the number of divisions n=0;
[0038] S3-2, the number of divisions n plus 1, that is, n = n + 1, calculate the average value m of the current HP, classify the housing price sample points greater than or equal to m as the head housing price points, and obtain the set H n ; Classify the housing price sample points less than m as tail housing price points, and get the set T n ;
[0039] S3-3. Calculate H n The sample proportion p′ is:
[0040]
[0041] If p′ is less than or equal to p=0.5, then the tail house price point T n Defined as the initial secondary market HT for the current split number n , and execute S3-4 and S3-5; otherwise, HP is defined as the initial secondary market HT of the current split number n , and execute S3-4 and S3-6.
[0042] S3-4. Initial secondary market HT n The housing price sample points are constructed into a network according to spatial information (i.e., longitude and latitude). The weight of the network edge is the Euclidean distance between the housing price sample points. The network edges with edge weight greater than r = 800 are deleted, and the scale of each subgraph in the network is calculated. The scale of each subgraph in the network is calculated, and the housing price sample points with a subgraph scale less than 3 are deleted to obtain the new secondary market HT n ′.
[0043] S3-5. Calculate HT n If s′ is greater than or equal to s=200, then HT n ′ is defined as the natural secondary market NHT n , execute S3-7; otherwise, HT n Add H n As the new HT n Execute S3-4 and S3-6.
[0044] S3-6. Calculate HT n If s′ is greater than or equal to s=200, then HT n ′ is defined as the natural secondary market NHTn , execute S3-8; otherwise, let n = n-1, and then divide T with n times of division n With HT n Add together to get the new HT n , then re-execute S3-4 and S3-6 until s′ is greater than or equal to s=200, and then execute S3-8.
[0045] S3-7, NHT n As the new HP, return to S3-2 and enter the next cycle.
[0046] S3-8, the cycle ends, and all natural secondary markets NHT are output n .
[0047] S4. The obtained natural market NHT n The hedonic price model was established using variables including house structure information, location characteristics, and facility distance. House structure information includes area, number of bedrooms, number of bathrooms, floors, etc. in data collection. Location characteristics include road selectivity and integration calculated using the spatial syntax software Depthmap. Facility distance includes the distance to hospitals, primary schools, middle schools, supermarkets, and other facilities.
[0048] S5. Call the statsmodels library in Python to use the variance_inflation_factor function to calculate multicollinearity for all variables, and gradually eliminate the factors with the highest VIF (variance inflation factor) and calculate VIF again until the VIF of all variables is less than or equal to 5, and obtain the variables for constructing the final model.
[0049] S6. Call the statsmodels library in Python and use the OLS function to perform multiple linear regression to obtain the NHT of different natural markets in each city. n The fitting results of .
[0050] S7. For each city in the Yangtze River Delta, the NHT of the adjacent natural markets after merging is relative to the individual natural markets. n The overall R 2 Increment size, get the final real estate secondary market segmentation results. The specific steps are as follows:
[0051] S7-1. The city's natural market NHT n The number is recorded as n;
[0052] S7-2, starting from the two-by-two merger, select 2 to n adjacent and continuous natural markets NHT n, merge the corresponding housing price sample points and establish a characteristic price model for multiple linear regression fitting to obtain the corresponding regression fit goodness of fit
[0053] S7-3. Calculate the relative value of each natural market combination to the NHT alone. n The increase in the goodness of fit of the multiple linear regression Select the maximum increment The natural market merger situation is that the merged secondary market is output as Obtain the final real estate secondary market segmentation results.
[0054] S8. Call the GeoPandas and Matplotlib libraries in Python to perform spatial visualization of the NS results obtained for 26 cities in the Yangtze River Delta.
[0055] like Figure 2 As shown in the figure, under the set parameters, the NS structure of each city in the Yangtze River Delta is obtained according to this method. The most levels are 4, such as Nanjing and Hefei, and the least is 1, such as Ma'anshan. This example shows the division of real estate sub-markets in the Yangtze River Delta region based on fractal structure.
[0056] Although the embodiments of the present invention have been disclosed as above, they are not limited to the applications listed in the specification and implementation modes, and can be fully applicable to various fields suitable for the present invention. For those familiar with the art, for those of ordinary skill in the art, various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principle and spirit of the present invention. Therefore, without departing from the general concept defined by the claims and the scope of equivalents, the present invention is not limited to specific details.
Claims
1. A method for identifying the structure of the real estate secondary market based on fractal structure, characterized in that: The following steps are involved: S1. Collect spatial vector data of housing prices in the main built-up areas of the research city, define them as housing price sample points, and obtain a set HP; the housing price sample points include: housing unit price information, housing structure information and housing latitude and longitude information; the housing structure information includes: area, number of bedrooms, number of bathrooms and floors; S2. Execute the head-to-tail segmentation loop on HP. The specific steps are as follows: S2-1. Set the values of the main parameters: number of divisions n, head-to-tail division ratio p, housing price sample point contact distance threshold r, housing price point sample size s; S2-2, let the number of divisions n = n + 1, and calculate the average value m of the current HP; classify the housing price sample points greater than or equal to m as the head housing price points, and obtain the set H n ; Classify the housing price sample points less than m as tail housing price points, and get the set T n ; S2-3. Calculate H n The sample proportion p′ is: S2-4. If p′ is less than or equal to p, then T n Defined as the initial secondary market HT for the current split number n , and execute S2-5 to S2-7; otherwise, HP is defined as the initial secondary market HT of the current split number n , and execute S2-5, S2-6 and S2-8; S2-5. Initial secondary market HT n The housing price sample points are networked according to the house longitude and latitude information, and the weight of the network edge is the Euclidean distance between the housing price sample points; S2-6. Delete the network edges with weight greater than r, calculate the size of each subgraph in the network, delete the housing price sample points with subgraph size less than 3, and obtain the new secondary market HT n ′; S2-7. Calculate HT n ′, the number of sample points of housing prices s′; if s′ is greater than or equal to s, then HT n ′ is defined as the natural secondary market NHT n , execute S2-9; otherwise, HT n Add the head H n As the new HT n , execute S2-5, S2-6 and S2-8; S2-8. Calculate HT n ′, the number of sample points of housing prices s′; if s′ is greater than or equal to s, then HT n ′ is defined as the natural secondary market NHT n , execute S2-10; otherwise, let n = n-1, and then divide T with n times of division n With HT n Add together to get the new HT n , then re-execute S2-5, S2-6 and S2-8 until s′ is greater than or equal to s, then execute S2-10; S2-9, NHT n As the new HP, return to S2-2 and enter the next cycle; S2-10, the cycle ends and all natural secondary market NHT is output n ; S3. The obtained natural market NHT n Establish a hedonic price model, using variables including: housing structure information, location characteristics and facility distance; use multiple linear regression to obtain different natural market NHT n The fitting results are as follows: According to the adjacent natural market after merging relative to the single natural market NHT n The overall R 2 Increment size to obtain the final real estate sub-market segmentation results.
2. The method for identifying the structure of the real estate secondary market based on fractal structure according to claim 1, characterized in that: The specific steps of S3 are as follows: S3-1. Natural Market NHT n The number of is denoted as k; S3-2, starting from the two-by-two merger, select 2 to k adjacent and continuous natural markets NHT n , merge the corresponding housing price sample points and establish a characteristic price model for multiple linear regression fitting to obtain the corresponding regression fit goodness of fit S3-3. Calculate the relative value of each natural market combination to the NHT alone. n The increase in the goodness of fit of the multiple linear regression Select the maximum increment Max The natural market merger situation, the merged secondary market output as NS n , and obtain the final real estate secondary market segmentation results.
3. The method for identifying the structure of the real estate secondary market based on fractal structure according to claim 1, characterized in that: The number of segmentation times n=0, the head-to-tail segmentation ratio p=0.5, the housing price point connection distance threshold r=800, and the housing price point sample size s=200.
Citation Information
Patent Citations
A real estate market monitoring and early warning method and system based on price index
CN109409951A
Real estate price intelligent evaluation method and device based on a multi-level evaluation model
CN112418936A
